Alloy

Minimal, OTP-native agent loop for Elixir.

Alloy is the completion-tool-call loop and nothing else. Send messages to any LLM, execute tool calls, loop until done. Swap providers with one line. Run agents as supervised GenServers. No opinions on sessions, persistence, memory, scheduling, or UI — those belong in your application.

{:ok, result} = Alloy.run("Read mix.exs and tell me the version",
  provider: {Alloy.Provider.OpenAI, api_key: System.get_env("OPENAI_API_KEY"), model: "gpt-5.4"},
  tools: [Alloy.Tool.Core.Read]
)

result.text #=> "The version is 0.9.0"

Why Alloy?

Most agent frameworks try to be everything — sessions, memory, RAG, multi-agent orchestration, scheduling, UI. Alloy does one thing well: the agent loop. Inspired by Pi Agent's minimalism, Alloy brings the same philosophy to the BEAM with OTP's natural advantages: supervision, fault isolation, parallel tool execution, and real concurrency.

3 providers — Anthropic, OpenAI, and OpenAICompat (works with any OpenAI-compatible API: Ollama, OpenRouter, xAI, DeepSeek, Mistral, Groq, Together, etc.)
4 built-in tools — read, write, edit, bash
GenServer agents — supervised, stateful, message-passing
Streaming — token-by-token from any provider, unified interface
Async dispatch — send_message/2 fires non-blocking, result arrives via PubSub
Middleware — custom hooks, tool blocking
Context compaction — summary-based compaction when approaching token limits, with configurable reserve and fallback to truncation
Prompt caching — Anthropic cache: true adds cache breakpoints for 60-90% input token savings
Reasoning blocks — DeepSeek/xAI reasoning_content parsed as first-class thinking blocks
Provider passthrough — extra_body injects arbitrary provider-specific params (response_format, temperature, reasoning_effort)
Telemetry — run, turn, provider, and compaction lifecycle events for OTEL/logging/metrics
Cost guard — max_budget_cents halts the loop before overspending
OTP-native — supervision trees, hot code reloading, real parallel tool execution
~5,000 lines — small enough to read, understand, and extend

Design Boundary

Alloy stays minimal by owning protocol and loop concerns, not application workflows.

What belongs in Alloy:

Provider wire-format translation
Tool-call / completion loop mechanics
Normalized message blocks
Opaque provider-owned state such as stored response IDs
Provider response metadata such as citations or server-side tool telemetry

What does not belong in Alloy:

Sessions and persistence policy
File storage, indexing, or retrieval workflows
UI rendering for citations, search, or artifacts
Scheduling, background job orchestration, or dashboards
Tenant plans, quotas, billing, or hosted infrastructure policy

Rule of thumb: if the feature is required to speak a provider API correctly, and could help any Alloy consumer, it likely belongs here. If it needs a database table, product defaults, UI decisions, or tenancy logic, it belongs in your application layer.

Installation

Add alloy to your dependencies in mix.exs:

def deps do
  [
    {:alloy, "~> 0.9"}
  ]
end

Quick Start

Simple completion

{:ok, result} = Alloy.run("What is 2+2?",
  provider: {Alloy.Provider.Anthropic, api_key: "sk-ant-...", model: "claude-sonnet-4-6"}
)

result.text #=> "4"

Agent with tools

{:ok, result} = Alloy.run("Read mix.exs and summarize the dependencies",
  provider: {Alloy.Provider.OpenAICompat,
    api_url: "https://generativelanguage.googleapis.com",
    chat_path: "/v1beta/openai/chat/completions",
    api_key: "...", model: "gemini-2.5-flash-lite"},
  tools: [Alloy.Tool.Core.Read, Alloy.Tool.Core.Bash],
  max_turns: 10
)

Gemini model IDs Alloy now budgets for include gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite, gemini-3-pro-preview, and gemini-3-flash-preview.

Swap providers in one line

# The same tools and conversation work with any provider
opts = [tools: [Alloy.Tool.Core.Read], max_turns: 10]

# Anthropic
Alloy.run("Read mix.exs", [{:provider, {Alloy.Provider.Anthropic, api_key: "...", model: "claude-sonnet-4-6"}} | opts])

# OpenAI
Alloy.run("Read mix.exs", [{:provider, {Alloy.Provider.OpenAI, api_key: "...", model: "gpt-5.4"}} | opts])

# xAI via Responses-compatible API
Alloy.run("Read mix.exs", [{:provider, {Alloy.Provider.OpenAI, api_key: "...", api_url: "https://api.x.ai", model: "grok-4"}} | opts])

# xAI via chat completions (reasoning models, extra_body)
Alloy.run("Read mix.exs", [{:provider, {Alloy.Provider.OpenAICompat, api_key: "...", api_url: "https://api.x.ai", model: "grok-4.1-fast-reasoning"}} | opts])

# Any OpenAI-compatible API (Ollama, OpenRouter, DeepSeek, Mistral, Groq, etc.)
Alloy.run("Read mix.exs", [{:provider, {Alloy.Provider.OpenAICompat, api_url: "http://localhost:11434", model: "llama4"}} | opts])

Streaming

For a one-shot run, use Alloy.stream/3:

{:ok, result} =
  Alloy.stream("Explain OTP", fn chunk ->
    IO.write(chunk)
  end,
    provider: {Alloy.Provider.OpenAI, api_key: "...", model: "gpt-5.4"}
  )

For a persistent agent process with conversation state, use Alloy.Agent.Server.stream_chat/4:

{:ok, agent} = Alloy.Agent.Server.start_link(
  provider: {Alloy.Provider.OpenAI, api_key: "...", model: "gpt-5.4"},
  tools: [Alloy.Tool.Core.Read]
)

{:ok, result} = Alloy.Agent.Server.stream_chat(agent, "Explain OTP", fn chunk ->
  IO.write(chunk)  # Print each token as it arrives
end)

All providers support streaming. If a custom provider doesn't implement stream/4, the turn loop falls back to complete/3 automatically.

Alloy.run/2 remains the buffered convenience API. Use Alloy.stream/3 when you want the same one-shot flow with token streaming.

Provider-owned state

Some provider APIs expose server-side state such as stored response IDs. That transport concern lives in Alloy; your app decides whether and how to persist it.

Results expose provider-owned state in result.metadata.provider_state:

{:ok, result} =
  Alloy.run("Read the repo",
    provider: {Alloy.Provider.OpenAI,
      api_key: System.get_env("XAI_API_KEY"),
      api_url: "https://api.x.ai",
      model: "grok-4",
      store: true
    }
  )

provider_state = result.metadata.provider_state

Pass that state back to the same provider on the next turn to continue a provider-native conversation:

{:ok, next_result} =
  Alloy.run("Keep going",
    messages: result.messages,
    provider: {Alloy.Provider.OpenAI,
      api_key: System.get_env("XAI_API_KEY"),
      api_url: "https://api.x.ai",
      model: "grok-4",
      provider_state: provider_state
    }
  )

Provider-native tools and citations

Responses-compatible providers can expose built-in server-side tools without leaking those wire details into your app layer.

For xAI search tools:

{:ok, result} =
  Alloy.run("Summarise the latest xAI docs updates",
    provider: {Alloy.Provider.OpenAI,
      api_key: System.get_env("XAI_API_KEY"),
      api_url: "https://api.x.ai",
      model: "grok-4",
      web_search: %{allowed_domains: ["docs.x.ai"]},
      include: ["inline_citations"]
    }
  )

Citation metadata is exposed in two places:

result.metadata.provider_response.citations for provider-level citation data
assistant text blocks may include :annotations for inline citation spans

Overriding model metadata

Alloy derives the compaction budget from the configured provider model when it knows that model's context window. If you need to support a just-released model before Alloy ships a catalog update, override it in config:

{:ok, result} = Alloy.run("Summarise this repository",
  provider: {Alloy.Provider.OpenAI, api_key: "...", model: "gpt-5.4-2026-03-05"},
  model_metadata_overrides: %{
    "gpt-5.4" => 900_000,
    "acme-reasoner" => %{limit: 640_000, suffix_patterns: ["", ~r/^-\d{4}\.\d{2}$/]}
  }
)

Set max_tokens explicitly when you want a fixed compaction budget. Otherwise Alloy derives it from the current model, including after Alloy.Agent.Server.set_model/2 switches to a different provider model.

Use compaction: when you want to tune how much room Alloy reserves before it summarizes older context:

{:ok, result} = Alloy.run("Summarise this repository",
  provider: {Alloy.Provider.OpenAI, api_key: "...", model: "gpt-5.4"},
  compaction: [
    reserve_tokens: 12_000,
    keep_recent_tokens: 8_000,
    fallback: :truncate
  ]
)

Cost guard

Cap how much an agent run can spend:

{:ok, result} = Alloy.run("Research this codebase thoroughly",
  provider: {Alloy.Provider.Anthropic, api_key: "...", model: "claude-sonnet-4-6"},
  tools: [Alloy.Tool.Core.Read, Alloy.Tool.Core.Bash],
  max_budget_cents: 50
)

case result.status do
  :completed -> IO.puts(result.text)
  :budget_exceeded -> IO.puts("Stopped: spent #{result.usage.estimated_cost_cents}¢")
end

Set max_budget_cents: nil (default) for no limit.

Anthropic prompt caching

Enable prompt caching to save 60-90% on input tokens. Alloy automatically adds cache_control breakpoints to the system prompt and last tool definition:

{:ok, result} = Alloy.run("Explain this codebase",
  provider: {Alloy.Provider.Anthropic,
    api_key: "...", model: "claude-sonnet-4-6",
    cache: true
  },
  tools: [Alloy.Tool.Core.Read, Alloy.Tool.Core.Bash],
  system_prompt: "You are a senior Elixir developer."
)

# Cache usage is reported in result.usage
result.usage.cache_creation_input_tokens  #=> 1500
result.usage.cache_read_input_tokens      #=> 1500  (on subsequent calls)

Reasoning model support (DeepSeek, xAI)

OpenAI-compatible reasoning models that return reasoning_content (DeepSeek-R1, xAI Grok reasoning variants) are automatically parsed into thinking blocks:

{:ok, result} = Alloy.run("Solve this step by step",
  provider: {Alloy.Provider.OpenAICompat,
    api_url: "https://api.x.ai",
    api_key: "...", model: "grok-4.1-fast-reasoning"
  }
)

# Thinking blocks are preserved in message content
[thinking, text] = hd(result.messages).content
thinking.type     #=> "thinking"
thinking.thinking #=> "Step 1: Let me consider..."
text.type         #=> "text"
text.text         #=> "The answer is 42."

Provider-specific parameters (extra_body)

Pass arbitrary provider-specific parameters via extra_body. It merges last, so it can override any default field:

{:ok, result} = Alloy.run("Return JSON",
  provider: {Alloy.Provider.OpenAICompat,
    api_url: "https://api.deepseek.com",
    api_key: "...", model: "deepseek-chat",
    extra_body: %{
      "response_format" => %{"type" => "json_object"},
      "temperature" => 0.3
    }
  }
)

Works for any provider param: reasoning_effort, max_completion_tokens, presence_penalty, etc.

Telemetry

Alloy emits telemetry events for observability. Attach handlers for OTEL, logging, or custom metrics:

:telemetry.attach_many("my-handler", [
  [:alloy, :run, :start],
  [:alloy, :run, :stop],
  [:alloy, :turn, :start],
  [:alloy, :turn, :stop],
  [:alloy, :provider, :request],
  [:alloy, :compaction, :done],
  [:alloy, :tool, :start],
  [:alloy, :tool, :stop],
  [:alloy, :event]
], &MyApp.Telemetry.handle_event/4, nil)

Event	Measurements	Metadata
`[:alloy, :run, :start]`	`system_time`	`model`
`[:alloy, :run, :stop]`	`duration_ms`	`status`, `turns`, `model`
`[:alloy, :turn, :start]`	`system_time`	`turn`
`[:alloy, :turn, :stop]`	—	`turn`, `status`
`[:alloy, :provider, :request]`	`duration_ms`	`provider`, `model`, `streaming`, `attempt`, `result`
`[:alloy, :compaction, :done]`	`messages_before`, `messages_after`	`turn`
`[:alloy, :tool, :start]`	—	tool identity, correlation
`[:alloy, :tool, :stop]`	`duration_ms`	tool identity, result

Supervised GenServer agent

{:ok, agent} = Alloy.Agent.Server.start_link(
  provider: {Alloy.Provider.Anthropic, api_key: "...", model: "claude-sonnet-4-6"},
  tools: [Alloy.Tool.Core.Read, Alloy.Tool.Core.Edit, Alloy.Tool.Core.Bash],
  system_prompt: "You are a senior Elixir developer."
)

{:ok, response} = Alloy.Agent.Server.chat(agent, "What does this project do?")
{:ok, response} = Alloy.Agent.Server.chat(agent, "Now refactor the main module")

Async dispatch (Phoenix LiveView)

Fire a message without blocking the caller — ideal for LiveView and background jobs:

# Subscribe to receive the result
Phoenix.PubSub.subscribe(MyApp.PubSub, "agent:#{session_id}:responses")

# Returns {:ok, request_id} immediately — agent works in the background
{:ok, req_id} = Alloy.Agent.Server.send_message(agent, "Summarise this report",
  request_id: "req-123"
)

# Handle the result whenever it arrives
def handle_info({:agent_response, %{text: text, request_id: "req-123"}}, socket) do
  {:noreply, assign(socket, :response, text)}
end

Providers

Vendor	Recommended Module	Example Models
Anthropic	`Alloy.Provider.Anthropic`	`claude-opus-4-6`, `claude-sonnet-4-6`, `claude-haiku-4-5`
OpenAI	`Alloy.Provider.OpenAI`	`gpt-5.4`
xAI	`Alloy.Provider.OpenAI` with `api_url: "https://api.x.ai"`	`grok-4`, `grok-4.1-fast`, `grok-4-fast-reasoning`, `grok-code-fast-1`
Gemini	`Alloy.Provider.OpenAICompat`	`gemini-2.5-pro`, `gemini-2.5-flash`, `gemini-2.5-flash-lite`, `gemini-3-pro-preview`
Other OpenAI-compatible APIs	`Alloy.Provider.OpenAICompat`	Ollama, OpenRouter, DeepSeek, Mistral, Groq, Together

Use Alloy.Provider.OpenAI for native Responses APIs like OpenAI and xAI. Use Alloy.Provider.OpenAICompat for chat-completions compatible APIs and local runtimes.

OpenAICompat works with any API that implements the OpenAI chat completions format. Just set api_url, model, and optionally api_key and chat_path.

Built-in Tools

Tool	Module	Description
read	`Alloy.Tool.Core.Read`	Read files from disk
write	`Alloy.Tool.Core.Write`	Write files to disk
edit	`Alloy.Tool.Core.Edit`	Search-and-replace editing
bash	`Alloy.Tool.Core.Bash`	Execute shell commands (restricted shell by default)

Custom tools

defmodule MyApp.Tools.WebSearch do
  @behaviour Alloy.Tool

  @impl true
  def name, do: "web_search"

  @impl true
  def description, do: "Search the web for information"

  @impl true
  def input_schema do
    %{
      type: "object",
      properties: %{query: %{type: "string", description: "Search query"}},
      required: ["query"]
    }
  end

  @impl true
  def execute(%{"query" => query}, _context) do
    # Your implementation here
    {:ok, "Results for: #{query}"}
  end
end

Code execution (Anthropic)

Enable Anthropic's server-side code execution sandbox:

{:ok, result} = Alloy.run("Calculate the first 20 Fibonacci numbers",
  provider: {Alloy.Provider.Anthropic, api_key: "...", model: "claude-sonnet-4-6"},
  code_execution: true
)

Architecture

Alloy.run/2                    One-shot agent loop (pure function)
Alloy.Agent.Server             GenServer wrapper (stateful, supervisable)
Alloy.Agent.Turn               Single turn: call provider → execute tools → return
Alloy.Provider                 Behaviour: translate wire format ↔ Alloy.Message
Alloy.Tool                     Behaviour: name, description, input_schema, execute
Alloy.Middleware               Pipeline: custom hooks, tool blocking
Alloy.Context.Compactor        Automatic conversation summarization

Sessions, persistence, multi-agent coordination, scheduling, skills, and UI belong in your application layer. See Anvil for a reference Phoenix application built on Alloy.

License

MIT — see LICENSE.

Releases

Hex.pm publishing is handled by GitHub Actions on v* tags. Successful publishes also dispatch the landing-site version sync workflow.

Name		Name	Last commit message	Last commit date
Latest commit History 73 Commits
.alloy/context		.alloy/context
.github/workflows		.github/workflows
lib		lib
test		test
.credo.exs		.credo.exs
.formatter.exs		.formatter.exs
.gitignore		.gitignore
CHANGELOG.md		CHANGELOG.md
CONTRIBUTING.md		CONTRIBUTING.md
LICENSE		LICENSE
README.md		README.md
mix.exs		mix.exs
mix.lock		mix.lock

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

Alloy

Why Alloy?

Design Boundary

Installation

Quick Start

Simple completion

Agent with tools

Swap providers in one line

Streaming

Provider-owned state

Provider-native tools and citations

Overriding model metadata

Cost guard

Anthropic prompt caching

Reasoning model support (DeepSeek, xAI)

Provider-specific parameters (extra_body)

Telemetry

Supervised GenServer agent

Async dispatch (Phoenix LiveView)

Providers

Built-in Tools

Custom tools

Code execution (Anthropic)

Architecture

License

Releases

About

Uh oh!

Releases 11

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

Alloy

Why Alloy?

Design Boundary

Installation

Quick Start

Simple completion

Agent with tools

Swap providers in one line

Streaming

Provider-owned state

Provider-native tools and citations

Overriding model metadata

Cost guard

Anthropic prompt caching

Reasoning model support (DeepSeek, xAI)

Provider-specific parameters (extra_body)

Telemetry

Supervised GenServer agent

Async dispatch (Phoenix LiveView)

Providers

Built-in Tools

Custom tools

Code execution (Anthropic)

Architecture

License

Releases

About

Topics

Resources

License

Contributing

Uh oh!

Stars

Watchers

Forks

Releases 11

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages